Using of metaheuristics for optimizing satellite image registration

The work presented in this paper focuses on the use of optimization techniques for the satellite images registration (SIR). The idea was reduction in treatment execution time and improving the quality of results. Our contribution is mainly in the introduction of the concept for interesting points in an image (characteristic points) which, once extracted, will be associated with an estimator similarity type SSD (sum of squared differences) for the assessment of the digital image registration operation. In addition to the above work, we did the mathematical modelling of the images registration problem, as well as we adapted and implemented the resolution method to the image registration problem. Our main contribution enables us to; adapt two optimization metaheuristics, namely simulated annealing and genetic algorithms, in their sequential and parallel versions; and the substantial reduction in the complexity of processing of image registration. This work can be extended to other fields like, the medical imaging, territory surveillance, the environmental protection etc.